Updates: Generative AI solutions: GPT-4o available now in all regions & enhancements to Data Transporter
➡️ Exact delivery dates may vary, and brands may therefore not have immediate access to all features on the date of publication. Please contact your LivePerson account team for the exact dates on which you will have access to the features.
🚨The timing and scope of these features or functionalities remain at the sole discretion of LivePerson and are subject to change.
Generative AI solutions: GPT-4o available now in all regions
If your solution uses Generative AI, and you’re using the LLMs made available via LivePerson (not your own LLM), this announcement is for you.
LivePerson is pleased to communicate that GPT-4o is now available in all regions, and we will be auto-upgrading all applications using GPT-4 on June 5, per Microsoft’s GPT-4 model retirement date of June 6, 2025.
For several months now, LivePerson has been using GPT-4o in Early Access products, and we’re very pleased with its performance.
If you are using a Routing AI agent or aiStudio Flows in Early Access, you will want to review the performance of your solutions after the upgrade to GPT-4o June 5, 2025. You also have the option to upgrade in advance through aiStudio, if you prefer. If you need support, we can help! Please reach out to your LivePerson team for support.
Our current models and where to use them
Our latest models are GPT-4o and GPT-4o mini.
- GPT-4o mini is the replacement model for GPT-3.5 turbo. GPT-4o mini powers our Knowledge AI Agents and Copilot suite (Automated Conversation Summaries, Copilot Rewrite, and Conversation Assist).
- GPT-4o is the replacement model for GPT-4. GPT-4o powers our Routing AI Agent, as well as Early Access products in aiStudio, such as the Data Collection Agent.
Why use different models for different products? In the course of developing and testing our products, LivePerson’s AI and Machine Learning team has determined that GPT-4o mini (and its predecessor, GPT-3.5 turbo) was not performant for some of our more complex use cases. While GPT-4o mini has improved on the earlier model, the Routing AI agent and response generation for our Data Collection flow in aiStudio both work better with GPT-4o.
Why upgrade to GPT-4o now? Microsoft Azure, LivePerson's partner for LLM models, has issued a retirement date of June 6, 2025 for GPT-4. In anticipation of this upgrade, earlier this year we made the GPT-4o model available in Early Access for some brands. And as per this release note, it is now available in all regions. The great news is that GPT-4o performs better than GPT-4 in our applications and testing.
Report Center Enhancements
Adds report selectors to Data Transporter export integration
This update introduces report selector options to the Data Transporter export integration. By default, flat file exports include the pre-configured fields from Data Transporter. However, users can now further customize exports via Report Center by selecting additional fields not included in the default configuration.
New Capabilities:
- Content to Retrieve (Dropdown): Available for Conversations and Messages flat files, allowing users to add extra fields.
- Dynamic Checkbox Options: Based on the selected flat file, users can configure additional settings (e.g., include transcripts, filter by PCS conversations).
- Tooltips: Click the Information icon next to each option for detailed descriptions.
These enhancements provide more flexibility and control over exported data formats.
Adds History to the Filters Dropdown to address the incident in Report Center
This update automatically saves the user's last selected filter values by default. To give users greater visibility and control over these saved selections, we've introduced a new "History" section.
Key Features:
- The History section displays the recently saved filter values.
- If previously selected values are no longer relevant or visible (e.g., due to changes in the selected timeframe), users can navigate to the History tab to view and deselect them.
- This ensures users can remove outdated or unavailable filter values that may not appear in the All section due to current data constraints.
This improves the user experience by making it easier to manage and adjust filter preferences.
Data Transporter
Real-Time API Permission Validation, add button to manually validate API keys
LivePerson has added a new validation step during task setup in Data Transporter to improve task reliability and reduce silent failures. Until now, users could configure tasks using API keys that lacked necessary permissions for selected reports. This resulted with subsequent task failures, with limited visibility into the root cause depending on notification setup and access to target email account.
Key Enhancements:
- Permission Check at Setup: When a user creates a task that depends on specific APIs (e.g., PDC, MI), Data Transporter will now automatically test selected “Data Source” (API key) to confirm it has all necessary permissions before allowing the task to be saved.
- Immediate Feedback: If the selected API keys do not have required permissions, an informative error message is displayed—blocking setup and guiding the user to resolution.
- Improved Visibility: This ensures permission issues are caught early and transparently, avoiding backend failures and unnecessary support escalations.
- Manual Trigger: The test can also be forced from the UI before attempting to save the task. Users can trigger the test immediately after selecting “Data Source” and report, by clicking the new “Test Report API” button. This shortens the feedback loop even further and speeds up selection of the right “Data Source”.
Benefits:
- Prevents task failures before they occur
- Eliminates reliance on backend error notifications
- Provides fast, clear feedback at the point of setup
- Reduces troubleshooting and support overhead
Expose more API parameters in selectors
LivePerson has expanded the number of API parameters exposed in the Data Transporter (DT) UI, making it easier for users to configure more precise and performant report queries.
Previously, many advanced API filtering options were not exposed in the DT UI, limiting flexibility and requiring manual workarounds. This enhancement aligns the UI with the extended capabilities of the underlying data APIs.
Key Enhancements
- Direct Exposure of API Parameters:
Users can now access additional API filtering parameters directly from the DT interface (e.g., PDC, MI). - No DT-Based Filtering Logic:
All filtering is handled by the API methods themselves—meaning it creates a source of truth for filtering definitions between the API and Data Transporter Reports. - Enhanced Usability Without Adding Reports:
Rather than introducing new reports, all parameters are added to existing reports, preserving report structure and minimizing user learning curves.
Benefits
- Greater query precision without engineering support
- Streamlined user experience—no new reports to learn
- Improved maintainability by reusing report logic
- Alignments between API definitions and Data Transporter Reports.
User Type Filter for PDC Report (Report ID: 11223)
To improve signal-to-noise ratio in exported reports, LivePerson has added a filter to the PDC report that allows users to select message types based on sender identity (Bot / Human / System).
Previously, the PDC report included messages from all participant types, resulting in high volumes of bot-generated data that made it difficult to focus on human agent interactions.
Key Enhancements
- New Filter Selector: A dropdown selector has been added to the PDC report configuration, allowing users to include or exclude messages by participant type:
- Human (userTypeId = 1)
- Bot (userTypeId = 2)
- System (userTypeId = 0)
- Improved Clarity in Exports: By filtering out bots or system messages, users can now tailor reports to specific analysis goals—e.g., QA of human agent responses, or tracking automation coverage.
Benefits
- Reduces report noise by excluding irrelevant participants
- Supports automation and human performance analysis independently
- Enhances report relevance and usability for downstream analysis